AI-BASED PAVEMENT CONDITION ASSESSMENT AND AUTOMATED REPORTING USING AN OFFLINE ANDROID APPLICATION

Abstract

Pavement condition surveys are essential for infrastructure maintenance planning, yet they are still conducted manually in many regions. Conventional practice relies on vehicle odometers for distance measurement, visual identification of surface distress, and separate GPS devicecoordinate logging, followed by manual transcription of the data into spreadsheets. This workflow is time-consuming and prone to human error. It also delays decision-making and reduces the effectiveness of maintenance prioritization. This paper presents an AI-based mobile application for pavement condition surveying and reporting that integrates data acquisition, storage, and reporting into a single Android platform. The system uses the device’s embedded GPS module for real-time location tracking and road-length measurement. Road surface images are captured through the onboard camera, and each image is automatically geotagged with latitude, longitude, and timestamp metadata. The system employs a to detect and classify pavement defects such as , achieving .All survey data are stored in local (offline) storage, which allows the system to operate in remote areas with limited or no network connectivity. On completion of a survey, the collected records can be exported to Excel formats with a single action, which eliminates manual data entry. The proposed system reduces survey time and human error, and it improves the accuracy, reliability, and cost-effectiveness of pavement data collection for maintenance engineers and supervisors.